1 LSST: a Complementary Probe of Dark Energy ¡ J. A. Tyson and D. M. Wittman , J. F. Hennawi and D. N. Spergel Bell Labs, Lucent Technologies, 600 Mountain Ave., Murray Hill, NJ 07974 ¡ Astrophysical Sciences, Princeton University, Princeton, NJ 08544 The number of mass clusters and their distribution in redshift are very sensitive to the density of matter ¢¤£ and the equation of state of dark energy ¥ . Using weak lens gravitational tomography one can detect clusters of dark matter, weigh them, £§¦ ¢ image their projected mass distribution, and determine their 3-D location. The degeneracy curve in the ¥ plane is nearly orthogonal to that from CMB or SN measurements. Thus, a combination of CMB data with weak lens tomography of clusters ¢¤£ can yield precision measurements of and ¥ , independently of the SN observations. The Large Synoptic Survey Telescope ¨ © © ( ¨ © © ) will repeatedly survey 30,000 square degrees of the sky in multiple wavelengths. will create a 3-D tomographic assay of mass overdensities back to half the age of the universe by measuring the shear and color-redshift of billions of high redshift galaxies. By simultaneously measuring several functions of cosmic shear and mass cluster abundance, ¨© © will provide a number of independent constraints on the dark energy density and the equation of state. 1. INTRODUCTION unbiased to baryons and radiation. will mea- sure the volume-redshift relation and the mass struc- To date, most of what we know about the large- ture development over the range of cosmic time dur- scale structure of the universe comes from the ob- ing which it is currently thought that the universe served anisotropies in the cosmic microwave back- transitioned from matter-dominated to dark energy- ground (CMB) and from the distribution of galaxies dominated. and supernovae. The CMB provides the earliest sam- The evolution of mass clustering is the most sen- ple of mass fluctuations, from a time when the uni- sitive test of our current dark energy and dark mat- verse was 50,000 times younger. Different cosmolog- ter cosmology[10,14]. Calibrated maps of mass as a ical models predict different volume-redshift relations function of cosmic look-back time can (1) constrain and different scenarios in the growth of mass struc- the nature of the dark matter by its power spectrum, tures over cosmic time, so comparison of the CMB- by the way it clumps gravitationally over time, and derived mass spectrum with that seen at later times by its detailed distribution (voids, walls, filaments); will be a powerful test of cosmology. The large-scale (2) test the cosmology through cosmic shear, the cu- mass distribution at late times has traditionally been mulative shear due to all mass overdensities out to characterized through the large-scale galaxy distribu- high redshift; (3) probe dark energy content in a way tion, on the assumption that galaxies trace mass in a complementary to and entirely independently from simple way. But weak gravitational lensing can trace CMB+SN, through the time evolution of the power structure more directly, is sensitive to the comoving spectrum; and (4) sharply constrain the dark energy volume, and relies on no “standard metrics.” For the equation of state (to about one percent) through the supernova test of the luminosity distance, there re- time evolution of the number of mass overdensities. main questions of whether Type Ia supernovae are ac- Such maps must be obtained from statistical inversion curate “standard candles” over the relevant range of of the observed shear of a billion high redshift back- look-back times. ¤ ground galaxies, but with existing facilities it would will utilize the physics-based technique of take hundreds of years to accumulate sufficient data gravitational lensing in which angles and redshifts are to definitively address these questions. will measured, yielding direct maps of dark mass in 3-D, address these questions within a decade. 2 2. WIDE DEEP FAST edly, other rare phenomena remain to be discovered as well. The various direct observational tests of cosmol- The design of the is driven by the following ogy (weak lensing, CMB anisotropy, and SNe) each figure of merit. In a given integration time, the size of suffer from degeneracies. The CMB measurements field larger than that can be explored to given stel- are currently by far the most accurate of the three, $#&%(' lar magnitude is directly proportional to !" , but still are sensitive only to a combination of dark where A is the collecting area, the solid angle of matter and dark energy rather than one or the other. # the field of view, the overall efficiency and d ) the The luminosity distance vs redshift (SNe Ia test) by solid angle of the seeing-limited image. Today’s 8m itself has moderate sensitivity to the equation of state class telescopes and detectors are superb at optimiz- of dark energy, but with an prior, SNe Ia observa- ing all of these factors except ) (typically is tions may be used to constrain the ratio of density to 0.04 deg ). pressure in the dark energy. Weak lensing measure- Advances in three areas of technology (large as- ments measure dark matter in a direct way. Each of pherical optics fabrication and metrology, semicon- these experiments has its strengths, and in combina- ductors and terascale computation, and ULSI CCD tion they break the degeneracies. But a set of exper- mosaics) have come together in the design of the iments which minimally breaks the degeneracies will system [12]. With its large throughput and never test the foundations. dedicated observational mode, the ¤ opens an The degeneracy can also be broken by MAP unexplored region of parameter space and enables CMB data, and even more cleanly later with the programs that would take many decades on current Planck data. The combination of weak lensing (with facilities. Previously named the “Dark Matter Tele- photometric redshifts) and CMB data provides a sen- ¤ scope,” the will be able to reach 5 * limiting sitive consistency test for the theory [11]. In addition surface brightness of 28-30 magnitude arcsec +, in the the combination will constrain most of the many pa- wavelength range 0.3–1 - m over a 7 deg field in 3 rameters of current theories substantially better than nights of dark time. This opens new observational either alone. will discover many supernovae possibilities in low surface brightness wide-field sur- in both the time-domain search and the 1000 deg sur- veys. For point sources, the will reach 24th vey (because it requires many visits to each piece of magnitude at 5 * in only 10 seconds. Repeated imag- sky). Combining the Type Ia SN results with either ing of large areas of the sky in this mode will probe the weak lensing or CMB results can also lift degen- unprecedented volumes out to high redshift in a way eracies in space. Combining all three obser- which offers control of image aberration systematics vations (weak lensing with photometric redshifts, SN, crucial to weak lensing. CMB) will lead to even higher precision cosmology. The time required to complete the surveys de- Comparison of different approaches will reveal sys- scribed above is inversely proportional to optical tematics and lead to refinements. Most importantly, throughput: square meters of collecting area times the combination will test the entire foundation of the ¤ solid angle on the sky. The with an 8.4m pri- theory. will produce both weak-lens and su- mary and 7 deg per exposure will have a throughput pernova probes of dark energy with the same instru- 50 times larger than current 4m telescopes with their ment. large CCD mosaics. The three-mirror design will pro- In addition, the large field of view opens up the duce unprecedented image quality over the full 7 deg time domain in search of rare, high-energy phenom- field. A single 10s exposure will go to 24th magnitude ena. Currently, time-domain astrophysics is limited to (5 * ). The dark mass/energy survey would utilize deep projects that provide shallow coverage of 10 deg ( /.)0 mag) multi-color imaging over 30,000 deg (e.g. the MACHO project), or extremely shallow cov- with photometric redshifts for source galaxies out to erage of a larger area. With the ability to survey an z=3. entire hemisphere to 24th magnitude in a few dark ¤ A team has been working on science and de- nights, will open up new areas of parameter sign for two years (see http://lsst.org). While there are space. For example, faint optical bursts. Undoubt- some technology challenges, no show-stoppers have 3 is highest at small angular separation and drops for widely separated galaxies whose light bundles travel through completely different structures. Because the typical background galaxy has an in- trinsic ellipticity of roughly 30%, many thousands of source galaxies must be averaged together to detect this small signal. In addition, a large area of sky must be covered, because mass structures should span a few arcminutes to a degree at a typical mid-path distance of redshift 0.5. Driven by advances in detector technology (CCDs) and software, the first detections of cosmic shear came from four different groups almost simultane- ously [17,13,1,9]. This first generation of cosmic shear measurements indicates a low- universe. While this is no suprise, it should be emphasized that this a new method, completely independent of the methods traditionally used to measure . The current generation of cosmic shear surveys will provide more precise constraints, but are all severely limited by telescope time and the small !1) of the existing wide-field telescopes. The Deep Lens Sur- vey (http://dls.bell-labs.com), for example, is planned Figure 1.